Employing Polarimetric Interferometry Synthetic Aperture Radar (PolInSAR) data to inverse forest parameters is a hot topic in the research field of PolInSAR. The typical forest parameter inversion algorithm is the three-stage inversion algorithm based on Random Volume over Ground (RVoG) model. The errors of linear fitting and volume scattering correlation estimation are the major factors for parameter estimation accuracy. In this paper, straight line fitting employing the total least squares method is used to estimate the ground phase. Then, the Gamma function is applied as the line measure to adaptively estimate the volume scattering correlation. The improved three-stage inversion algorithm with PolInSAR is presented. The experiment result proves the forest parameters inversion result is accurate and reliable.